@alextis59/athena 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +15 -0
- package/README.md +147 -0
- package/dist/agent/ask.d.ts +7 -0
- package/dist/agent/ask.d.ts.map +1 -0
- package/dist/agent/ask.js +35 -0
- package/dist/agent/ask.js.map +1 -0
- package/dist/agent/citations.d.ts +27 -0
- package/dist/agent/citations.d.ts.map +1 -0
- package/dist/agent/citations.js +190 -0
- package/dist/agent/citations.js.map +1 -0
- package/dist/agent/fake-provider.d.ts +9 -0
- package/dist/agent/fake-provider.d.ts.map +1 -0
- package/dist/agent/fake-provider.js +51 -0
- package/dist/agent/fake-provider.js.map +1 -0
- package/dist/agent/index.d.ts +10 -0
- package/dist/agent/index.d.ts.map +1 -0
- package/dist/agent/index.js +6 -0
- package/dist/agent/index.js.map +1 -0
- package/dist/agent/loop.d.ts +11 -0
- package/dist/agent/loop.d.ts.map +1 -0
- package/dist/agent/loop.js +156 -0
- package/dist/agent/loop.js.map +1 -0
- package/dist/agent/providers.d.ts +17 -0
- package/dist/agent/providers.d.ts.map +1 -0
- package/dist/agent/providers.js +171 -0
- package/dist/agent/providers.js.map +1 -0
- package/dist/agent/tools.d.ts +8 -0
- package/dist/agent/tools.d.ts.map +1 -0
- package/dist/agent/tools.js +233 -0
- package/dist/agent/tools.js.map +1 -0
- package/dist/agent/types.d.ts +84 -0
- package/dist/agent/types.d.ts.map +1 -0
- package/dist/agent/types.js +2 -0
- package/dist/agent/types.js.map +1 -0
- package/dist/cli/help.d.ts +2 -0
- package/dist/cli/help.d.ts.map +1 -0
- package/dist/cli/help.js +28 -0
- package/dist/cli/help.js.map +1 -0
- package/dist/cli/main.d.ts +3 -0
- package/dist/cli/main.d.ts.map +1 -0
- package/dist/cli/main.js +5 -0
- package/dist/cli/main.js.map +1 -0
- package/dist/cli/options.d.ts +21 -0
- package/dist/cli/options.d.ts.map +1 -0
- package/dist/cli/options.js +97 -0
- package/dist/cli/options.js.map +1 -0
- package/dist/cli/run.d.ts +21 -0
- package/dist/cli/run.d.ts.map +1 -0
- package/dist/cli/run.js +144 -0
- package/dist/cli/run.js.map +1 -0
- package/dist/config/index.d.ts +4 -0
- package/dist/config/index.d.ts.map +1 -0
- package/dist/config/index.js +3 -0
- package/dist/config/index.js.map +1 -0
- package/dist/config/loader.d.ts +14 -0
- package/dist/config/loader.d.ts.map +1 -0
- package/dist/config/loader.js +352 -0
- package/dist/config/loader.js.map +1 -0
- package/dist/config/schema.d.ts +6 -0
- package/dist/config/schema.d.ts.map +1 -0
- package/dist/config/schema.js +72 -0
- package/dist/config/schema.js.map +1 -0
- package/dist/config/types.d.ts +53 -0
- package/dist/config/types.d.ts.map +1 -0
- package/dist/config/types.js +2 -0
- package/dist/config/types.js.map +1 -0
- package/dist/doctor/doctor.d.ts +47 -0
- package/dist/doctor/doctor.d.ts.map +1 -0
- package/dist/doctor/doctor.js +159 -0
- package/dist/doctor/doctor.js.map +1 -0
- package/dist/embeddings/index.d.ts +117 -0
- package/dist/embeddings/index.d.ts.map +1 -0
- package/dist/embeddings/index.js +515 -0
- package/dist/embeddings/index.js.map +1 -0
- package/dist/errors.d.ts +19 -0
- package/dist/errors.d.ts.map +1 -0
- package/dist/errors.js +37 -0
- package/dist/errors.js.map +1 -0
- package/dist/index.d.ts +17 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +9 -0
- package/dist/index.js.map +1 -0
- package/dist/indexer/discovery.d.ts +10 -0
- package/dist/indexer/discovery.d.ts.map +1 -0
- package/dist/indexer/discovery.js +38 -0
- package/dist/indexer/discovery.js.map +1 -0
- package/dist/indexer/hash.d.ts +3 -0
- package/dist/indexer/hash.d.ts.map +1 -0
- package/dist/indexer/hash.js +11 -0
- package/dist/indexer/hash.js.map +1 -0
- package/dist/indexer/index.d.ts +7 -0
- package/dist/indexer/index.d.ts.map +1 -0
- package/dist/indexer/index.js +5 -0
- package/dist/indexer/index.js.map +1 -0
- package/dist/indexer/kinds.d.ts +4 -0
- package/dist/indexer/kinds.d.ts.map +1 -0
- package/dist/indexer/kinds.js +38 -0
- package/dist/indexer/kinds.js.map +1 -0
- package/dist/indexer/run.d.ts +42 -0
- package/dist/indexer/run.d.ts.map +1 -0
- package/dist/indexer/run.js +563 -0
- package/dist/indexer/run.js.map +1 -0
- package/dist/logging/logger.d.ts +11 -0
- package/dist/logging/logger.d.ts.map +1 -0
- package/dist/logging/logger.js +18 -0
- package/dist/logging/logger.js.map +1 -0
- package/dist/mcp/index.d.ts +35 -0
- package/dist/mcp/index.d.ts.map +1 -0
- package/dist/mcp/index.js +392 -0
- package/dist/mcp/index.js.map +1 -0
- package/dist/parsers/chunking.d.ts +4 -0
- package/dist/parsers/chunking.d.ts.map +1 -0
- package/dist/parsers/chunking.js +112 -0
- package/dist/parsers/chunking.js.map +1 -0
- package/dist/parsers/index.d.ts +8 -0
- package/dist/parsers/index.d.ts.map +1 -0
- package/dist/parsers/index.js +5 -0
- package/dist/parsers/index.js.map +1 -0
- package/dist/parsers/line-parser.d.ts +13 -0
- package/dist/parsers/line-parser.d.ts.map +1 -0
- package/dist/parsers/line-parser.js +242 -0
- package/dist/parsers/line-parser.js.map +1 -0
- package/dist/parsers/normalize.d.ts +4 -0
- package/dist/parsers/normalize.d.ts.map +1 -0
- package/dist/parsers/normalize.js +17 -0
- package/dist/parsers/normalize.js.map +1 -0
- package/dist/parsers/parser.d.ts +14 -0
- package/dist/parsers/parser.d.ts.map +1 -0
- package/dist/parsers/parser.js +77 -0
- package/dist/parsers/parser.js.map +1 -0
- package/dist/parsers/pdf2llm.d.ts +14 -0
- package/dist/parsers/pdf2llm.d.ts.map +1 -0
- package/dist/parsers/pdf2llm.js +215 -0
- package/dist/parsers/pdf2llm.js.map +1 -0
- package/dist/parsers/types.d.ts +48 -0
- package/dist/parsers/types.d.ts.map +1 -0
- package/dist/parsers/types.js +2 -0
- package/dist/parsers/types.js.map +1 -0
- package/dist/retrieval/index.d.ts +4 -0
- package/dist/retrieval/index.d.ts.map +1 -0
- package/dist/retrieval/index.js +3 -0
- package/dist/retrieval/index.js.map +1 -0
- package/dist/retrieval/schemas.d.ts +127 -0
- package/dist/retrieval/schemas.d.ts.map +1 -0
- package/dist/retrieval/schemas.js +65 -0
- package/dist/retrieval/schemas.js.map +1 -0
- package/dist/retrieval/tools.d.ts +31 -0
- package/dist/retrieval/tools.d.ts.map +1 -0
- package/dist/retrieval/tools.js +803 -0
- package/dist/retrieval/tools.js.map +1 -0
- package/dist/retrieval/types.d.ts +104 -0
- package/dist/retrieval/types.d.ts.map +1 -0
- package/dist/retrieval/types.js +2 -0
- package/dist/retrieval/types.js.map +1 -0
- package/dist/server/index.d.ts +3 -0
- package/dist/server/index.d.ts.map +1 -0
- package/dist/server/index.js +2 -0
- package/dist/server/index.js.map +1 -0
- package/dist/server/server.d.ts +49 -0
- package/dist/server/server.d.ts.map +1 -0
- package/dist/server/server.js +517 -0
- package/dist/server/server.js.map +1 -0
- package/dist/shared/types.d.ts +98 -0
- package/dist/shared/types.d.ts.map +1 -0
- package/dist/shared/types.js +2 -0
- package/dist/shared/types.js.map +1 -0
- package/dist/storage/database.d.ts +18 -0
- package/dist/storage/database.d.ts.map +1 -0
- package/dist/storage/database.js +66 -0
- package/dist/storage/database.js.map +1 -0
- package/dist/storage/index.d.ts +8 -0
- package/dist/storage/index.d.ts.map +1 -0
- package/dist/storage/index.js +5 -0
- package/dist/storage/index.js.map +1 -0
- package/dist/storage/migrations.d.ts +15 -0
- package/dist/storage/migrations.d.ts.map +1 -0
- package/dist/storage/migrations.js +170 -0
- package/dist/storage/migrations.js.map +1 -0
- package/dist/storage/paths.d.ts +15 -0
- package/dist/storage/paths.d.ts.map +1 -0
- package/dist/storage/paths.js +27 -0
- package/dist/storage/paths.js.map +1 -0
- package/dist/storage/schema.d.ts +11 -0
- package/dist/storage/schema.d.ts.map +1 -0
- package/dist/storage/schema.js +40 -0
- package/dist/storage/schema.js.map +1 -0
- package/dist/ui/index.d.ts +2 -0
- package/dist/ui/index.d.ts.map +1 -0
- package/dist/ui/index.js +2 -0
- package/dist/ui/index.js.map +1 -0
- package/dist/ui/static-app.d.ts +2 -0
- package/dist/ui/static-app.d.ts.map +1 -0
- package/dist/ui/static-app.js +735 -0
- package/dist/ui/static-app.js.map +1 -0
- package/docs/adr/0001-sqlite-driver.md +23 -0
- package/docs/agentic-rag-specifications.md +659 -0
- package/docs/agentic-rag-study.md +1056 -0
- package/docs/implementation-plan.md +1425 -0
- package/docs/mcp-integration.md +154 -0
- package/docs/validation-corpus.md +71 -0
- package/package.json +79 -0
|
@@ -0,0 +1,242 @@
|
|
|
1
|
+
import { normalizeText, stripMarkdownHeadingDecorators } from "./normalize.js";
|
|
2
|
+
export function parseMarkdownLikeLines(sourceText, options) {
|
|
3
|
+
const lines = splitLines(sourceText);
|
|
4
|
+
const blocks = [];
|
|
5
|
+
const warnings = [];
|
|
6
|
+
const headingStack = [];
|
|
7
|
+
let paragraphLines = [];
|
|
8
|
+
let fenceStartLine = null;
|
|
9
|
+
let codeLines = [];
|
|
10
|
+
for (const [index, line] of lines.entries()) {
|
|
11
|
+
const lineNumber = index + 1;
|
|
12
|
+
const fenceMatch = line.match(/^```|^~~~~/u);
|
|
13
|
+
if (fenceStartLine !== null) {
|
|
14
|
+
codeLines.push({ line, lineNumber });
|
|
15
|
+
if (fenceMatch !== null) {
|
|
16
|
+
blocks.push(createBlock("code", codeLines, headingStack));
|
|
17
|
+
codeLines = [];
|
|
18
|
+
fenceStartLine = null;
|
|
19
|
+
}
|
|
20
|
+
continue;
|
|
21
|
+
}
|
|
22
|
+
if (fenceMatch !== null) {
|
|
23
|
+
flushParagraph();
|
|
24
|
+
fenceStartLine = lineNumber;
|
|
25
|
+
codeLines = [{ line, lineNumber }];
|
|
26
|
+
continue;
|
|
27
|
+
}
|
|
28
|
+
const headingMatch = line.match(/^(#{1,6})\s+(.+)$/u);
|
|
29
|
+
if (headingMatch !== null) {
|
|
30
|
+
flushParagraph();
|
|
31
|
+
const depth = headingMatch[1]?.length ?? 1;
|
|
32
|
+
const heading = stripMarkdownHeadingDecorators(headingMatch[2] ?? "");
|
|
33
|
+
headingStack.splice(depth - 1);
|
|
34
|
+
headingStack[depth - 1] = heading;
|
|
35
|
+
blocks.push(createBlock("heading", [{ line: heading, lineNumber }], headingStack.filter(Boolean)));
|
|
36
|
+
continue;
|
|
37
|
+
}
|
|
38
|
+
if (line.trim() === "") {
|
|
39
|
+
flushParagraph();
|
|
40
|
+
continue;
|
|
41
|
+
}
|
|
42
|
+
paragraphLines.push({ line, lineNumber });
|
|
43
|
+
}
|
|
44
|
+
if (fenceStartLine !== null) {
|
|
45
|
+
warnings.push({
|
|
46
|
+
code: options.mdx ? "mdx.unclosed_fence" : "markdown.unclosed_fence",
|
|
47
|
+
message: "Unclosed fenced code block.",
|
|
48
|
+
startLine: fenceStartLine,
|
|
49
|
+
endLine: lines.length
|
|
50
|
+
});
|
|
51
|
+
blocks.push(createBlock("code", codeLines, headingStack));
|
|
52
|
+
}
|
|
53
|
+
flushParagraph();
|
|
54
|
+
return { blocks, warnings };
|
|
55
|
+
function flushParagraph() {
|
|
56
|
+
if (paragraphLines.length === 0) {
|
|
57
|
+
return;
|
|
58
|
+
}
|
|
59
|
+
blocks.push(createBlock(classifyParagraph(paragraphLines), paragraphLines, headingStack));
|
|
60
|
+
paragraphLines = [];
|
|
61
|
+
}
|
|
62
|
+
}
|
|
63
|
+
export function parsePlainTextLines(sourceText) {
|
|
64
|
+
const lines = splitLines(sourceText);
|
|
65
|
+
const blocks = [];
|
|
66
|
+
let paragraphLines = [];
|
|
67
|
+
for (const [index, line] of lines.entries()) {
|
|
68
|
+
const lineNumber = index + 1;
|
|
69
|
+
if (line.trim() === "") {
|
|
70
|
+
flushParagraph();
|
|
71
|
+
continue;
|
|
72
|
+
}
|
|
73
|
+
paragraphLines.push({ line, lineNumber });
|
|
74
|
+
}
|
|
75
|
+
flushParagraph();
|
|
76
|
+
return { blocks, warnings: [] };
|
|
77
|
+
function flushParagraph() {
|
|
78
|
+
if (paragraphLines.length === 0) {
|
|
79
|
+
return;
|
|
80
|
+
}
|
|
81
|
+
blocks.push(createBlock("paragraph", paragraphLines, []));
|
|
82
|
+
paragraphLines = [];
|
|
83
|
+
}
|
|
84
|
+
}
|
|
85
|
+
export function parseRstLines(sourceText) {
|
|
86
|
+
const lines = splitLines(sourceText);
|
|
87
|
+
const blocks = [];
|
|
88
|
+
const headingStack = [];
|
|
89
|
+
let paragraphLines = [];
|
|
90
|
+
for (let index = 0; index < lines.length; index += 1) {
|
|
91
|
+
const line = lines[index] ?? "";
|
|
92
|
+
const nextLine = lines[index + 1] ?? "";
|
|
93
|
+
const lineNumber = index + 1;
|
|
94
|
+
if (line.trim() !== "" &&
|
|
95
|
+
/^([=\-~^"`:#*+])\1{2,}\s*$/u.test(nextLine.trim())) {
|
|
96
|
+
flushParagraph();
|
|
97
|
+
const depth = headingDepth(nextLine.trim()[0] ?? "=");
|
|
98
|
+
headingStack.splice(depth - 1);
|
|
99
|
+
headingStack[depth - 1] = line.trim();
|
|
100
|
+
blocks.push(createBlock("heading", [{ line: line.trim(), lineNumber }], headingStack.filter(Boolean)));
|
|
101
|
+
index += 1;
|
|
102
|
+
continue;
|
|
103
|
+
}
|
|
104
|
+
if (line.trim() === "") {
|
|
105
|
+
flushParagraph();
|
|
106
|
+
continue;
|
|
107
|
+
}
|
|
108
|
+
paragraphLines.push({ line, lineNumber });
|
|
109
|
+
}
|
|
110
|
+
flushParagraph();
|
|
111
|
+
return { blocks, warnings: [] };
|
|
112
|
+
function flushParagraph() {
|
|
113
|
+
if (paragraphLines.length === 0) {
|
|
114
|
+
return;
|
|
115
|
+
}
|
|
116
|
+
blocks.push(createBlock(classifyParagraph(paragraphLines), paragraphLines, headingStack));
|
|
117
|
+
paragraphLines = [];
|
|
118
|
+
}
|
|
119
|
+
}
|
|
120
|
+
export function parseAsciiDocLines(sourceText) {
|
|
121
|
+
const lines = splitLines(sourceText);
|
|
122
|
+
const blocks = [];
|
|
123
|
+
const headingStack = [];
|
|
124
|
+
let paragraphLines = [];
|
|
125
|
+
for (const [index, line] of lines.entries()) {
|
|
126
|
+
const lineNumber = index + 1;
|
|
127
|
+
const headingMatch = line.match(/^(={1,6})\s+(.+)$/u);
|
|
128
|
+
if (headingMatch !== null) {
|
|
129
|
+
flushParagraph();
|
|
130
|
+
const depth = headingMatch[1]?.length ?? 1;
|
|
131
|
+
const heading = headingMatch[2]?.trim() ?? "";
|
|
132
|
+
headingStack.splice(depth - 1);
|
|
133
|
+
headingStack[depth - 1] = heading;
|
|
134
|
+
blocks.push(createBlock("heading", [{ line: heading, lineNumber }], headingStack.filter(Boolean)));
|
|
135
|
+
continue;
|
|
136
|
+
}
|
|
137
|
+
if (line.trim() === "") {
|
|
138
|
+
flushParagraph();
|
|
139
|
+
continue;
|
|
140
|
+
}
|
|
141
|
+
paragraphLines.push({ line, lineNumber });
|
|
142
|
+
}
|
|
143
|
+
flushParagraph();
|
|
144
|
+
return { blocks, warnings: [] };
|
|
145
|
+
function flushParagraph() {
|
|
146
|
+
if (paragraphLines.length === 0) {
|
|
147
|
+
return;
|
|
148
|
+
}
|
|
149
|
+
blocks.push(createBlock(classifyParagraph(paragraphLines), paragraphLines, headingStack));
|
|
150
|
+
paragraphLines = [];
|
|
151
|
+
}
|
|
152
|
+
}
|
|
153
|
+
export function parseHtmlLines(sourceText) {
|
|
154
|
+
const lines = splitLines(sourceText);
|
|
155
|
+
const blocks = [];
|
|
156
|
+
const headingStack = [];
|
|
157
|
+
let ignoringRaw = false;
|
|
158
|
+
for (const [index, line] of lines.entries()) {
|
|
159
|
+
const lineNumber = index + 1;
|
|
160
|
+
const lower = line.toLowerCase();
|
|
161
|
+
if (/<(script|style)\b/u.test(lower)) {
|
|
162
|
+
ignoringRaw = true;
|
|
163
|
+
}
|
|
164
|
+
if (ignoringRaw) {
|
|
165
|
+
if (/<\/(script|style)>/u.test(lower)) {
|
|
166
|
+
ignoringRaw = false;
|
|
167
|
+
}
|
|
168
|
+
continue;
|
|
169
|
+
}
|
|
170
|
+
const headingMatch = line.match(/<h([1-6])\b[^>]*>(.*?)<\/h\1>/iu);
|
|
171
|
+
if (headingMatch !== null) {
|
|
172
|
+
const depth = Number.parseInt(headingMatch[1] ?? "1", 10);
|
|
173
|
+
const heading = decodeBasicEntities(stripTags(headingMatch[2] ?? "")).trim();
|
|
174
|
+
headingStack.splice(depth - 1);
|
|
175
|
+
headingStack[depth - 1] = heading;
|
|
176
|
+
blocks.push(createBlock("heading", [{ line: heading, lineNumber }], headingStack.filter(Boolean)));
|
|
177
|
+
continue;
|
|
178
|
+
}
|
|
179
|
+
const text = decodeBasicEntities(stripTags(line)).trim();
|
|
180
|
+
if (text !== "") {
|
|
181
|
+
blocks.push(createBlock(classifyHtmlLine(line), [{ line: text, lineNumber }], headingStack.filter(Boolean)));
|
|
182
|
+
}
|
|
183
|
+
}
|
|
184
|
+
return { blocks, warnings: [] };
|
|
185
|
+
}
|
|
186
|
+
function createBlock(kind, lines, headingPath) {
|
|
187
|
+
const text = lines
|
|
188
|
+
.map((line) => line.line)
|
|
189
|
+
.join("\n")
|
|
190
|
+
.trim();
|
|
191
|
+
return {
|
|
192
|
+
kind,
|
|
193
|
+
text,
|
|
194
|
+
normalizedText: normalizeText(text),
|
|
195
|
+
startLine: lines[0]?.lineNumber ?? 1,
|
|
196
|
+
endLine: lines.at(-1)?.lineNumber ?? lines[0]?.lineNumber ?? 1,
|
|
197
|
+
headingPath: [...headingPath]
|
|
198
|
+
};
|
|
199
|
+
}
|
|
200
|
+
function classifyParagraph(lines) {
|
|
201
|
+
const firstLine = lines[0]?.line.trim() ?? "";
|
|
202
|
+
if (/^(\*|-|\+|\d+\.)\s+/u.test(firstLine)) {
|
|
203
|
+
return "list";
|
|
204
|
+
}
|
|
205
|
+
if (firstLine.includes("|")) {
|
|
206
|
+
return "table";
|
|
207
|
+
}
|
|
208
|
+
return "paragraph";
|
|
209
|
+
}
|
|
210
|
+
function classifyHtmlLine(line) {
|
|
211
|
+
if (/<(li)\b/iu.test(line)) {
|
|
212
|
+
return "list";
|
|
213
|
+
}
|
|
214
|
+
if (/<(td|th|tr|table)\b/iu.test(line)) {
|
|
215
|
+
return "table";
|
|
216
|
+
}
|
|
217
|
+
if (/<(pre|code)\b/iu.test(line)) {
|
|
218
|
+
return "code";
|
|
219
|
+
}
|
|
220
|
+
return "paragraph";
|
|
221
|
+
}
|
|
222
|
+
function headingDepth(marker) {
|
|
223
|
+
const order = ["=", "-", "~", "^", '"', "`", ":", "#", "*", "+"];
|
|
224
|
+
const index = order.indexOf(marker);
|
|
225
|
+
return index === -1 ? 1 : index + 1;
|
|
226
|
+
}
|
|
227
|
+
function splitLines(text) {
|
|
228
|
+
return text.replace(/\r\n?/gu, "\n").split("\n");
|
|
229
|
+
}
|
|
230
|
+
function stripTags(text) {
|
|
231
|
+
return text.replace(/<[^>]*>/gu, " ");
|
|
232
|
+
}
|
|
233
|
+
function decodeBasicEntities(text) {
|
|
234
|
+
return text
|
|
235
|
+
.replace(/ /gu, " ")
|
|
236
|
+
.replace(/&/gu, "&")
|
|
237
|
+
.replace(/</gu, "<")
|
|
238
|
+
.replace(/>/gu, ">")
|
|
239
|
+
.replace(/"/gu, '"')
|
|
240
|
+
.replace(/'/gu, "'");
|
|
241
|
+
}
|
|
242
|
+
//# sourceMappingURL=line-parser.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"line-parser.js","sourceRoot":"","sources":["../../src/parsers/line-parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,8BAA8B,EAAE,MAAM,gBAAgB,CAAC;AAQ/E,MAAM,UAAU,sBAAsB,CACpC,UAAkB,EAClB,OAAyB;IAEzB,MAAM,KAAK,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACrC,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,MAAM,QAAQ,GAAoB,EAAE,CAAC;IACrC,MAAM,YAAY,GAAa,EAAE,CAAC;IAClC,IAAI,cAAc,GAA2C,EAAE,CAAC;IAChE,IAAI,cAAc,GAAkB,IAAI,CAAC;IACzC,IAAI,SAAS,GAA2C,EAAE,CAAC;IAE3D,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;QAC5C,MAAM,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;QAC7B,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,CAAC;QAE7C,IAAI,cAAc,KAAK,IAAI,EAAE,CAAC;YAC5B,SAAS,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;YAErC,IAAI,UAAU,KAAK,IAAI,EAAE,CAAC;gBACxB,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,SAAS,EAAE,YAAY,CAAC,CAAC,CAAC;gBAC1D,SAAS,GAAG,EAAE,CAAC;gBACf,cAAc,GAAG,IAAI,CAAC;YACxB,CAAC;YAED,SAAS;QACX,CAAC;QAED,IAAI,UAAU,KAAK,IAAI,EAAE,CAAC;YACxB,cAAc,EAAE,CAAC;YACjB,cAAc,GAAG,UAAU,CAAC;YAC5B,SAAS,GAAG,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;YACnC,SAAS;QACX,CAAC;QAED,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;QAEtD,IAAI,YAAY,KAAK,IAAI,EAAE,CAAC;YAC1B,cAAc,EAAE,CAAC;YACjB,MAAM,KAAK,GAAG,YAAY,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC;YAC3C,MAAM,OAAO,GAAG,8BAA8B,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YACtE,YAAY,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YAC/B,YAAY,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,OAAO,CAAC;YAClC,MAAM,CAAC,IAAI,CACT,WAAW,CACT,SAAS,EACT,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC,EAC/B,YAAY,CAAC,MAAM,CAAC,OAAO,CAAC,CAC7B,CACF,CAAC;YACF,SAAS;QACX,CAAC;QAED,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YACvB,cAAc,EAAE,CAAC;YACjB,SAAS;QACX,CAAC;QAED,cAAc,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;IAC5C,CAAC;IAED,IAAI,cAAc,KAAK,IAAI,EAAE,CAAC;QAC5B,QAAQ,CAAC,IAAI,CAAC;YACZ,IAAI,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,oBAAoB,CAAC,CAAC,CAAC,yBAAyB;YACpE,OAAO,EAAE,6BAA6B;YACtC,SAAS,EAAE,cAAc;YACzB,OAAO,EAAE,KAAK,CAAC,MAAM;SACtB,CAAC,CAAC;QACH,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,SAAS,EAAE,YAAY,CAAC,CAAC,CAAC;IAC5D,CAAC;IAED,cAAc,EAAE,CAAC;IAEjB,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC;IAE5B,SAAS,cAAc;QACrB,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAChC,OAAO;QACT,CAAC;QAED,MAAM,CAAC,IAAI,CACT,WAAW,CACT,iBAAiB,CAAC,cAAc,CAAC,EACjC,cAAc,EACd,YAAY,CACb,CACF,CAAC;QACF,cAAc,GAAG,EAAE,CAAC;IACtB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,UAAkB;IACpD,MAAM,KAAK,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACrC,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,IAAI,cAAc,GAA2C,EAAE,CAAC;IAEhE,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;QAC5C,MAAM,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;QAE7B,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YACvB,cAAc,EAAE,CAAC;YACjB,SAAS;QACX,CAAC;QAED,cAAc,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;IAC5C,CAAC;IAED,cAAc,EAAE,CAAC;IAEjB,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IAEhC,SAAS,cAAc;QACrB,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAChC,OAAO;QACT,CAAC;QAED,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,WAAW,EAAE,cAAc,EAAE,EAAE,CAAC,CAAC,CAAC;QAC1D,cAAc,GAAG,EAAE,CAAC;IACtB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,UAAkB;IAC9C,MAAM,KAAK,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACrC,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,MAAM,YAAY,GAAa,EAAE,CAAC;IAClC,IAAI,cAAc,GAA2C,EAAE,CAAC;IAEhE,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACrD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAChC,MAAM,QAAQ,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QACxC,MAAM,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;QAE7B,IACE,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE;YAClB,6BAA6B,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,EACnD,CAAC;YACD,cAAc,EAAE,CAAC;YACjB,MAAM,KAAK,GAAG,YAAY,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC;YACtD,YAAY,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YAC/B,YAAY,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;YACtC,MAAM,CAAC,IAAI,CACT,WAAW,CACT,SAAS,EACT,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,EAAE,UAAU,EAAE,CAAC,EACnC,YAAY,CAAC,MAAM,CAAC,OAAO,CAAC,CAC7B,CACF,CAAC;YACF,KAAK,IAAI,CAAC,CAAC;YACX,SAAS;QACX,CAAC;QAED,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YACvB,cAAc,EAAE,CAAC;YACjB,SAAS;QACX,CAAC;QAED,cAAc,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;IAC5C,CAAC;IAED,cAAc,EAAE,CAAC;IACjB,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IAEhC,SAAS,cAAc;QACrB,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAChC,OAAO;QACT,CAAC;QAED,MAAM,CAAC,IAAI,CACT,WAAW,CACT,iBAAiB,CAAC,cAAc,CAAC,EACjC,cAAc,EACd,YAAY,CACb,CACF,CAAC;QACF,cAAc,GAAG,EAAE,CAAC;IACtB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,UAAkB;IACnD,MAAM,KAAK,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACrC,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,MAAM,YAAY,GAAa,EAAE,CAAC;IAClC,IAAI,cAAc,GAA2C,EAAE,CAAC;IAEhE,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;QAC5C,MAAM,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;QAC7B,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;QAEtD,IAAI,YAAY,KAAK,IAAI,EAAE,CAAC;YAC1B,cAAc,EAAE,CAAC;YACjB,MAAM,KAAK,GAAG,YAAY,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC;YAC3C,MAAM,OAAO,GAAG,YAAY,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;YAC9C,YAAY,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YAC/B,YAAY,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,OAAO,CAAC;YAClC,MAAM,CAAC,IAAI,CACT,WAAW,CACT,SAAS,EACT,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC,EAC/B,YAAY,CAAC,MAAM,CAAC,OAAO,CAAC,CAC7B,CACF,CAAC;YACF,SAAS;QACX,CAAC;QAED,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YACvB,cAAc,EAAE,CAAC;YACjB,SAAS;QACX,CAAC;QAED,cAAc,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;IAC5C,CAAC;IAED,cAAc,EAAE,CAAC;IACjB,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IAEhC,SAAS,cAAc;QACrB,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAChC,OAAO;QACT,CAAC;QAED,MAAM,CAAC,IAAI,CACT,WAAW,CACT,iBAAiB,CAAC,cAAc,CAAC,EACjC,cAAc,EACd,YAAY,CACb,CACF,CAAC;QACF,cAAc,GAAG,EAAE,CAAC;IACtB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,UAAkB;IAC/C,MAAM,KAAK,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACrC,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,MAAM,YAAY,GAAa,EAAE,CAAC;IAClC,IAAI,WAAW,GAAG,KAAK,CAAC;IAExB,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;QAC5C,MAAM,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;QAC7B,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QAEjC,IAAI,oBAAoB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;YACrC,WAAW,GAAG,IAAI,CAAC;QACrB,CAAC;QAED,IAAI,WAAW,EAAE,CAAC;YAChB,IAAI,qBAAqB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtC,WAAW,GAAG,KAAK,CAAC;YACtB,CAAC;YACD,SAAS;QACX,CAAC;QAED,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,iCAAiC,CAAC,CAAC;QAEnE,IAAI,YAAY,KAAK,IAAI,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC;YAC1D,MAAM,OAAO,GAAG,mBAAmB,CACjC,SAAS,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CACjC,CAAC,IAAI,EAAE,CAAC;YACT,YAAY,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YAC/B,YAAY,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,OAAO,CAAC;YAClC,MAAM,CAAC,IAAI,CACT,WAAW,CACT,SAAS,EACT,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC,EAC/B,YAAY,CAAC,MAAM,CAAC,OAAO,CAAC,CAC7B,CACF,CAAC;YACF,SAAS;QACX,CAAC;QAED,MAAM,IAAI,GAAG,mBAAmB,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAEzD,IAAI,IAAI,KAAK,EAAE,EAAE,CAAC;YAChB,MAAM,CAAC,IAAI,CACT,WAAW,CACT,gBAAgB,CAAC,IAAI,CAAC,EACtB,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,EAC5B,YAAY,CAAC,MAAM,CAAC,OAAO,CAAC,CAC7B,CACF,CAAC;QACJ,CAAC;IACH,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AAClC,CAAC;AAED,SAAS,WAAW,CAClB,IAAyB,EACzB,KAA6C,EAC7C,WAAqB;IAErB,MAAM,IAAI,GAAG,KAAK;SACf,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;SACxB,IAAI,CAAC,IAAI,CAAC;SACV,IAAI,EAAE,CAAC;IAEV,OAAO;QACL,IAAI;QACJ,IAAI;QACJ,cAAc,EAAE,aAAa,CAAC,IAAI,CAAC;QACnC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,UAAU,IAAI,CAAC;QACpC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,UAAU,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,UAAU,IAAI,CAAC;QAC9D,WAAW,EAAE,CAAC,GAAG,WAAW,CAAC;KAC9B,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CACxB,KAA6C;IAE7C,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC;IAE9C,IAAI,sBAAsB,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QAC3C,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,OAAO,CAAC;IACjB,CAAC;IAED,OAAO,WAAW,CAAC;AACrB,CAAC;AAED,SAAS,gBAAgB,CAAC,IAAY;IACpC,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC3B,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,IAAI,uBAAuB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACvC,OAAO,OAAO,CAAC;IACjB,CAAC;IAED,IAAI,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACjC,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,OAAO,WAAW,CAAC;AACrB,CAAC;AAED,SAAS,YAAY,CAAC,MAAc;IAClC,MAAM,KAAK,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC;IACjE,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IAEpC,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;AACtC,CAAC;AAED,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,IAAI,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;AACnD,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC;AACxC,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,OAAO,IAAI;SACR,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC;SACxB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC;SACxB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC;AAC7B,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"normalize.d.ts","sourceRoot":"","sources":["../../src/parsers/normalize.ts"],"names":[],"mappings":"AAAA,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAElD;AAED,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAQnD;AAED,wBAAgB,8BAA8B,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAKnE"}
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
export function normalizeText(text) {
|
|
2
|
+
return text.replace(/\s+/g, " ").trim().toLowerCase();
|
|
3
|
+
}
|
|
4
|
+
export function estimateTokens(text) {
|
|
5
|
+
const trimmed = text.trim();
|
|
6
|
+
if (trimmed === "") {
|
|
7
|
+
return 0;
|
|
8
|
+
}
|
|
9
|
+
return trimmed.split(/\s+/).length;
|
|
10
|
+
}
|
|
11
|
+
export function stripMarkdownHeadingDecorators(text) {
|
|
12
|
+
return text
|
|
13
|
+
.replace(/\s*\{\/\*.*?\*\/\}\s*$/u, "")
|
|
14
|
+
.replace(/\s+#+\s*$/u, "")
|
|
15
|
+
.trim();
|
|
16
|
+
}
|
|
17
|
+
//# sourceMappingURL=normalize.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"normalize.js","sourceRoot":"","sources":["../../src/parsers/normalize.ts"],"names":[],"mappings":"AAAA,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;AACxD,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,IAAY;IACzC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAE5B,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;QACnB,OAAO,CAAC,CAAC;IACX,CAAC;IAED,OAAO,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC;AACrC,CAAC;AAED,MAAM,UAAU,8BAA8B,CAAC,IAAY;IACzD,OAAO,IAAI;SACR,OAAO,CAAC,yBAAyB,EAAE,EAAE,CAAC;SACtC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC;SACzB,IAAI,EAAE,CAAC;AACZ,CAAC"}
|
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
import { type Pdf2LlmConverter } from "./pdf2llm.js";
|
|
2
|
+
import type { PdfConfig } from "../config/index.js";
|
|
3
|
+
import type { ParsedDocument } from "./types.js";
|
|
4
|
+
import type { SourceKind } from "../shared/types.js";
|
|
5
|
+
export declare const TEXT_PARSER_VERSION = "1";
|
|
6
|
+
export interface ParseSourceFileOptions {
|
|
7
|
+
pdf: PdfConfig;
|
|
8
|
+
pdfConverter?: Pdf2LlmConverter;
|
|
9
|
+
}
|
|
10
|
+
export declare function parseSourceFile(absolutePath: string, sourceKind: SourceKind, options: ParseSourceFileOptions): Promise<ParsedDocument | null>;
|
|
11
|
+
export declare function parseSourceText(sourceText: string, sourceKind: SourceKind): Pick<ParsedDocument, "blocks" | "warnings">;
|
|
12
|
+
export declare function parserNameForKind(sourceKind: SourceKind): string;
|
|
13
|
+
export declare function parserVersionForKind(sourceKind: SourceKind): string;
|
|
14
|
+
//# sourceMappingURL=parser.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parser.d.ts","sourceRoot":"","sources":["../../src/parsers/parser.ts"],"names":[],"mappings":"AAUA,OAAO,EAIL,KAAK,gBAAgB,EACtB,MAAM,cAAc,CAAC;AACtB,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AACpD,OAAO,KAAK,EAAe,cAAc,EAAE,MAAM,YAAY,CAAC;AAC9D,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAErD,eAAO,MAAM,mBAAmB,MAAM,CAAC;AAEvC,MAAM,WAAW,sBAAsB;IACrC,GAAG,EAAE,SAAS,CAAC;IACf,YAAY,CAAC,EAAE,gBAAgB,CAAC;CACjC;AAED,wBAAsB,eAAe,CACnC,YAAY,EAAE,MAAM,EACpB,UAAU,EAAE,UAAU,EACtB,OAAO,EAAE,sBAAsB,GAC9B,OAAO,CAAC,cAAc,GAAG,IAAI,CAAC,CA4BhC;AAED,wBAAgB,eAAe,CAC7B,UAAU,EAAE,MAAM,EAClB,UAAU,EAAE,UAAU,GACrB,IAAI,CAAC,cAAc,EAAE,QAAQ,GAAG,UAAU,CAAC,CAkB7C;AAED,wBAAgB,iBAAiB,CAAC,UAAU,EAAE,UAAU,GAAG,MAAM,CAmBhE;AAED,wBAAgB,oBAAoB,CAAC,UAAU,EAAE,UAAU,GAAG,MAAM,CAEnE"}
|
|
@@ -0,0 +1,77 @@
|
|
|
1
|
+
import { readFile } from "node:fs/promises";
|
|
2
|
+
import { parseAsciiDocLines, parseHtmlLines, parseMarkdownLikeLines, parsePlainTextLines, parseRstLines } from "./line-parser.js";
|
|
3
|
+
import { normalizeText } from "./normalize.js";
|
|
4
|
+
import { PDF2LLM_PARSER_NAME, PDF2LLM_PARSER_VERSION, parsePdfFile } from "./pdf2llm.js";
|
|
5
|
+
export const TEXT_PARSER_VERSION = "1";
|
|
6
|
+
export async function parseSourceFile(absolutePath, sourceKind, options) {
|
|
7
|
+
if (sourceKind === "pdf") {
|
|
8
|
+
return await parsePdfFile(absolutePath, {
|
|
9
|
+
config: options.pdf,
|
|
10
|
+
...(options.pdfConverter === undefined
|
|
11
|
+
? {}
|
|
12
|
+
: { converter: options.pdfConverter })
|
|
13
|
+
});
|
|
14
|
+
}
|
|
15
|
+
if (sourceKind === "unknown") {
|
|
16
|
+
return null;
|
|
17
|
+
}
|
|
18
|
+
const sourceText = await readFile(absolutePath, "utf8");
|
|
19
|
+
const parsed = parseSourceText(sourceText, sourceKind);
|
|
20
|
+
const text = parsed.blocks.map((block) => block.text).join("\n\n");
|
|
21
|
+
return {
|
|
22
|
+
parserName: parserNameForKind(sourceKind),
|
|
23
|
+
parserVersion: TEXT_PARSER_VERSION,
|
|
24
|
+
sourceKind,
|
|
25
|
+
title: firstHeading(parsed.blocks),
|
|
26
|
+
text,
|
|
27
|
+
normalizedText: normalizeText(text),
|
|
28
|
+
blocks: parsed.blocks,
|
|
29
|
+
warnings: parsed.warnings
|
|
30
|
+
};
|
|
31
|
+
}
|
|
32
|
+
export function parseSourceText(sourceText, sourceKind) {
|
|
33
|
+
switch (sourceKind) {
|
|
34
|
+
case "markdown":
|
|
35
|
+
return parseMarkdownLikeLines(sourceText, { mdx: false });
|
|
36
|
+
case "mdx":
|
|
37
|
+
return parseMarkdownLikeLines(sourceText, { mdx: true });
|
|
38
|
+
case "text":
|
|
39
|
+
return parsePlainTextLines(sourceText);
|
|
40
|
+
case "rst":
|
|
41
|
+
return parseRstLines(sourceText);
|
|
42
|
+
case "asciidoc":
|
|
43
|
+
return parseAsciiDocLines(sourceText);
|
|
44
|
+
case "html":
|
|
45
|
+
return parseHtmlLines(sourceText);
|
|
46
|
+
case "pdf":
|
|
47
|
+
case "unknown":
|
|
48
|
+
return { blocks: [], warnings: [] };
|
|
49
|
+
}
|
|
50
|
+
}
|
|
51
|
+
export function parserNameForKind(sourceKind) {
|
|
52
|
+
switch (sourceKind) {
|
|
53
|
+
case "markdown":
|
|
54
|
+
return "markdown-line-parser";
|
|
55
|
+
case "mdx":
|
|
56
|
+
return "mdx-line-parser";
|
|
57
|
+
case "text":
|
|
58
|
+
return "text-line-parser";
|
|
59
|
+
case "rst":
|
|
60
|
+
return "rst-line-parser";
|
|
61
|
+
case "asciidoc":
|
|
62
|
+
return "asciidoc-line-parser";
|
|
63
|
+
case "html":
|
|
64
|
+
return "html-line-parser";
|
|
65
|
+
case "pdf":
|
|
66
|
+
return PDF2LLM_PARSER_NAME;
|
|
67
|
+
case "unknown":
|
|
68
|
+
return "unknown";
|
|
69
|
+
}
|
|
70
|
+
}
|
|
71
|
+
export function parserVersionForKind(sourceKind) {
|
|
72
|
+
return sourceKind === "pdf" ? PDF2LLM_PARSER_VERSION : TEXT_PARSER_VERSION;
|
|
73
|
+
}
|
|
74
|
+
function firstHeading(blocks) {
|
|
75
|
+
return blocks.find((block) => block.kind === "heading")?.text ?? null;
|
|
76
|
+
}
|
|
77
|
+
//# sourceMappingURL=parser.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parser.js","sourceRoot":"","sources":["../../src/parsers/parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAE5C,OAAO,EACL,kBAAkB,EAClB,cAAc,EACd,sBAAsB,EACtB,mBAAmB,EACnB,aAAa,EACd,MAAM,kBAAkB,CAAC;AAC1B,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EACL,mBAAmB,EACnB,sBAAsB,EACtB,YAAY,EAEb,MAAM,cAAc,CAAC;AAKtB,MAAM,CAAC,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAOvC,MAAM,CAAC,KAAK,UAAU,eAAe,CACnC,YAAoB,EACpB,UAAsB,EACtB,OAA+B;IAE/B,IAAI,UAAU,KAAK,KAAK,EAAE,CAAC;QACzB,OAAO,MAAM,YAAY,CAAC,YAAY,EAAE;YACtC,MAAM,EAAE,OAAO,CAAC,GAAG;YACnB,GAAG,CAAC,OAAO,CAAC,YAAY,KAAK,SAAS;gBACpC,CAAC,CAAC,EAAE;gBACJ,CAAC,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,YAAY,EAAE,CAAC;SACzC,CAAC,CAAC;IACL,CAAC;IAED,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,UAAU,GAAG,MAAM,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC;IACxD,MAAM,MAAM,GAAG,eAAe,CAAC,UAAU,EAAE,UAAU,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAEnE,OAAO;QACL,UAAU,EAAE,iBAAiB,CAAC,UAAU,CAAC;QACzC,aAAa,EAAE,mBAAmB;QAClC,UAAU;QACV,KAAK,EAAE,YAAY,CAAC,MAAM,CAAC,MAAM,CAAC;QAClC,IAAI;QACJ,cAAc,EAAE,aAAa,CAAC,IAAI,CAAC;QACnC,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;KAC1B,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,eAAe,CAC7B,UAAkB,EAClB,UAAsB;IAEtB,QAAQ,UAAU,EAAE,CAAC;QACnB,KAAK,UAAU;YACb,OAAO,sBAAsB,CAAC,UAAU,EAAE,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAC;QAC5D,KAAK,KAAK;YACR,OAAO,sBAAsB,CAAC,UAAU,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QAC3D,KAAK,MAAM;YACT,OAAO,mBAAmB,CAAC,UAAU,CAAC,CAAC;QACzC,KAAK,KAAK;YACR,OAAO,aAAa,CAAC,UAAU,CAAC,CAAC;QACnC,KAAK,UAAU;YACb,OAAO,kBAAkB,CAAC,UAAU,CAAC,CAAC;QACxC,KAAK,MAAM;YACT,OAAO,cAAc,CAAC,UAAU,CAAC,CAAC;QACpC,KAAK,KAAK,CAAC;QACX,KAAK,SAAS;YACZ,OAAO,EAAE,MAAM,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IACxC,CAAC;AACH,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,UAAsB;IACtD,QAAQ,UAAU,EAAE,CAAC;QACnB,KAAK,UAAU;YACb,OAAO,sBAAsB,CAAC;QAChC,KAAK,KAAK;YACR,OAAO,iBAAiB,CAAC;QAC3B,KAAK,MAAM;YACT,OAAO,kBAAkB,CAAC;QAC5B,KAAK,KAAK;YACR,OAAO,iBAAiB,CAAC;QAC3B,KAAK,UAAU;YACb,OAAO,sBAAsB,CAAC;QAChC,KAAK,MAAM;YACT,OAAO,kBAAkB,CAAC;QAC5B,KAAK,KAAK;YACR,OAAO,mBAAmB,CAAC;QAC7B,KAAK,SAAS;YACZ,OAAO,SAAS,CAAC;IACrB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAC,UAAsB;IACzD,OAAO,UAAU,KAAK,KAAK,CAAC,CAAC,CAAC,sBAAsB,CAAC,CAAC,CAAC,mBAAmB,CAAC;AAC7E,CAAC;AAED,SAAS,YAAY,CAAC,MAAqB;IACzC,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,EAAE,IAAI,IAAI,IAAI,CAAC;AACxE,CAAC"}
|
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
import { convertPdfToMarkdown } from "pdf-2-llm/node";
|
|
2
|
+
import type { PdfConfig } from "../config/index.js";
|
|
3
|
+
import type { ParsedDocument } from "./types.js";
|
|
4
|
+
export declare const PDF2LLM_PARSER_NAME = "pdf-2-llm";
|
|
5
|
+
export declare const PDF2LLM_PACKAGE_VERSION = "1.1.0";
|
|
6
|
+
export declare const PDF2LLM_PARSER_VERSION = "1.1.0";
|
|
7
|
+
export type Pdf2LlmConverter = typeof convertPdfToMarkdown;
|
|
8
|
+
export interface ParsePdfFileOptions {
|
|
9
|
+
config: PdfConfig;
|
|
10
|
+
converter?: Pdf2LlmConverter;
|
|
11
|
+
}
|
|
12
|
+
export declare function parsePdfFile(absolutePath: string, options: ParsePdfFileOptions): Promise<ParsedDocument>;
|
|
13
|
+
export declare function createPdfParserOptionsHash(config: PdfConfig): string;
|
|
14
|
+
//# sourceMappingURL=pdf2llm.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pdf2llm.d.ts","sourceRoot":"","sources":["../../src/parsers/pdf2llm.ts"],"names":[],"mappings":"AAEA,OAAO,EACL,oBAAoB,EAOrB,MAAM,gBAAgB,CAAC;AAKxB,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AACpD,OAAO,KAAK,EAEV,cAAc,EAGf,MAAM,YAAY,CAAC;AAEpB,eAAO,MAAM,mBAAmB,cAAc,CAAC;AAC/C,eAAO,MAAM,uBAAuB,UAAU,CAAC;AAC/C,eAAO,MAAM,sBAAsB,UAA0B,CAAC;AAE9D,MAAM,MAAM,gBAAgB,GAAG,OAAO,oBAAoB,CAAC;AAE3D,MAAM,WAAW,mBAAmB;IAClC,MAAM,EAAE,SAAS,CAAC;IAClB,SAAS,CAAC,EAAE,gBAAgB,CAAC;CAC9B;AAOD,wBAAsB,YAAY,CAChC,YAAY,EAAE,MAAM,EACpB,OAAO,EAAE,mBAAmB,GAC3B,OAAO,CAAC,cAAc,CAAC,CAmCzB;AAED,wBAAgB,0BAA0B,CAAC,MAAM,EAAE,SAAS,GAAG,MAAM,CASpE"}
|
|
@@ -0,0 +1,215 @@
|
|
|
1
|
+
import { createHash } from "node:crypto";
|
|
2
|
+
import { convertPdfToMarkdown, schemaVersion } from "pdf-2-llm/node";
|
|
3
|
+
import { UserFacingError } from "../errors.js";
|
|
4
|
+
import { parseMarkdownLikeLines } from "./line-parser.js";
|
|
5
|
+
import { normalizeText } from "./normalize.js";
|
|
6
|
+
export const PDF2LLM_PARSER_NAME = "pdf-2-llm";
|
|
7
|
+
export const PDF2LLM_PACKAGE_VERSION = "1.1.0";
|
|
8
|
+
export const PDF2LLM_PARSER_VERSION = PDF2LLM_PACKAGE_VERSION;
|
|
9
|
+
export async function parsePdfFile(absolutePath, options) {
|
|
10
|
+
const result = await convertWithTimeout(absolutePath, options);
|
|
11
|
+
const parsed = parseMarkdownLikeLines(result.markdown, { mdx: false });
|
|
12
|
+
const warnings = [
|
|
13
|
+
...parsed.warnings,
|
|
14
|
+
...toParserWarnings(result, options.config)
|
|
15
|
+
];
|
|
16
|
+
const blocks = annotatePdfBlocks(parsed.blocks, result, createLineStartOffsets(result.markdown), warnings);
|
|
17
|
+
const text = blocks.map((block) => block.text).join("\n\n");
|
|
18
|
+
return {
|
|
19
|
+
parserName: PDF2LLM_PARSER_NAME,
|
|
20
|
+
parserVersion: PDF2LLM_PARSER_VERSION,
|
|
21
|
+
sourceKind: "pdf",
|
|
22
|
+
title: firstHeading(blocks),
|
|
23
|
+
text,
|
|
24
|
+
normalizedText: normalizeText(text),
|
|
25
|
+
blocks,
|
|
26
|
+
warnings,
|
|
27
|
+
metadata: {
|
|
28
|
+
pdf2llm: {
|
|
29
|
+
packageVersion: PDF2LLM_PACKAGE_VERSION,
|
|
30
|
+
schemaVersion,
|
|
31
|
+
confidence: result.confidence,
|
|
32
|
+
pageCount: result.diagnostics.pages.length,
|
|
33
|
+
sourceType: result.ir.sourceType
|
|
34
|
+
}
|
|
35
|
+
},
|
|
36
|
+
artifacts: createPdfArtifacts(result, options.config)
|
|
37
|
+
};
|
|
38
|
+
}
|
|
39
|
+
export function createPdfParserOptionsHash(config) {
|
|
40
|
+
return hashString(JSON.stringify({
|
|
41
|
+
parser: PDF2LLM_PARSER_NAME,
|
|
42
|
+
packageVersion: PDF2LLM_PACKAGE_VERSION,
|
|
43
|
+
schemaVersion,
|
|
44
|
+
options: createPdfConvertOptionsSummary(config)
|
|
45
|
+
}));
|
|
46
|
+
}
|
|
47
|
+
function convertWithTimeout(absolutePath, options) {
|
|
48
|
+
const converter = options.converter ?? convertPdfToMarkdown;
|
|
49
|
+
const controller = new AbortController();
|
|
50
|
+
const timeout = setTimeout(() => {
|
|
51
|
+
controller.abort();
|
|
52
|
+
}, options.config.timeoutMs);
|
|
53
|
+
return converter(absolutePath, {
|
|
54
|
+
...createPdfConvertOptions(options.config),
|
|
55
|
+
signal: controller.signal
|
|
56
|
+
})
|
|
57
|
+
.catch((error) => {
|
|
58
|
+
if (controller.signal.aborted) {
|
|
59
|
+
throw new UserFacingError(`PDF parsing timed out after ${String(options.config.timeoutMs)} ms: ${absolutePath}`, error);
|
|
60
|
+
}
|
|
61
|
+
throw new UserFacingError(`PDF parsing failed: ${absolutePath}`, error);
|
|
62
|
+
})
|
|
63
|
+
.finally(() => {
|
|
64
|
+
clearTimeout(timeout);
|
|
65
|
+
});
|
|
66
|
+
}
|
|
67
|
+
function createPdfConvertOptions(config) {
|
|
68
|
+
return {
|
|
69
|
+
markdown: {
|
|
70
|
+
pageAnchors: config.pageAnchors
|
|
71
|
+
},
|
|
72
|
+
parser: {
|
|
73
|
+
mode: "tolerant"
|
|
74
|
+
},
|
|
75
|
+
security: {
|
|
76
|
+
timeoutMs: config.timeoutMs
|
|
77
|
+
},
|
|
78
|
+
tables: {
|
|
79
|
+
csvSidecars: config.csvSidecars
|
|
80
|
+
}
|
|
81
|
+
};
|
|
82
|
+
}
|
|
83
|
+
function createPdfConvertOptionsSummary(config) {
|
|
84
|
+
return {
|
|
85
|
+
markdown: {
|
|
86
|
+
pageAnchors: config.pageAnchors
|
|
87
|
+
},
|
|
88
|
+
parser: {
|
|
89
|
+
mode: "tolerant"
|
|
90
|
+
},
|
|
91
|
+
security: {
|
|
92
|
+
timeoutMs: config.timeoutMs
|
|
93
|
+
},
|
|
94
|
+
tables: {
|
|
95
|
+
csvSidecars: config.csvSidecars
|
|
96
|
+
},
|
|
97
|
+
citation: {
|
|
98
|
+
minConfidenceForAutoCitation: config.minConfidenceForAutoCitation
|
|
99
|
+
}
|
|
100
|
+
};
|
|
101
|
+
}
|
|
102
|
+
function annotatePdfBlocks(blocks, result, lineStartOffsets, warnings) {
|
|
103
|
+
const warningCodes = warnings.map((warning) => warning.code);
|
|
104
|
+
return blocks.map((block) => {
|
|
105
|
+
const match = findSourceMapMatch(result.sourceMap, lineStartOffsets[block.startLine - 1] ?? 0);
|
|
106
|
+
if (match === null) {
|
|
107
|
+
return {
|
|
108
|
+
...block,
|
|
109
|
+
metadata: {
|
|
110
|
+
...(block.metadata ?? {}),
|
|
111
|
+
pdfConfidence: result.confidence.overall,
|
|
112
|
+
pdfWarningCodes: warningCodes
|
|
113
|
+
}
|
|
114
|
+
};
|
|
115
|
+
}
|
|
116
|
+
return {
|
|
117
|
+
...block,
|
|
118
|
+
metadata: {
|
|
119
|
+
...(block.metadata ?? {}),
|
|
120
|
+
pdfPage: match.region.pageIndex + 1,
|
|
121
|
+
pdfConfidence: result.confidence.overall,
|
|
122
|
+
pdfWarningCodes: warningCodes,
|
|
123
|
+
pdfSourceMapKind: match.entry.kind,
|
|
124
|
+
pdfRegion: {
|
|
125
|
+
pageIndex: match.region.pageIndex,
|
|
126
|
+
x: match.region.x,
|
|
127
|
+
y: match.region.y,
|
|
128
|
+
width: match.region.width,
|
|
129
|
+
height: match.region.height,
|
|
130
|
+
source: match.region.source
|
|
131
|
+
}
|
|
132
|
+
}
|
|
133
|
+
};
|
|
134
|
+
});
|
|
135
|
+
}
|
|
136
|
+
function findSourceMapMatch(sourceMap, markdownOffset) {
|
|
137
|
+
for (const entry of sourceMap.entries) {
|
|
138
|
+
if (markdownOffset < entry.markdownStart ||
|
|
139
|
+
markdownOffset > entry.markdownEnd) {
|
|
140
|
+
continue;
|
|
141
|
+
}
|
|
142
|
+
const region = entry.regions[0];
|
|
143
|
+
if (region !== undefined) {
|
|
144
|
+
return { entry, region };
|
|
145
|
+
}
|
|
146
|
+
}
|
|
147
|
+
return null;
|
|
148
|
+
}
|
|
149
|
+
function createLineStartOffsets(markdown) {
|
|
150
|
+
const offsets = [0];
|
|
151
|
+
for (let index = 0; index < markdown.length; index += 1) {
|
|
152
|
+
if (markdown[index] === "\n") {
|
|
153
|
+
offsets.push(index + 1);
|
|
154
|
+
}
|
|
155
|
+
}
|
|
156
|
+
return offsets;
|
|
157
|
+
}
|
|
158
|
+
function toParserWarnings(result, config) {
|
|
159
|
+
const warnings = result.warnings.map((warning) => ({
|
|
160
|
+
code: warning.code,
|
|
161
|
+
message: warning.message
|
|
162
|
+
}));
|
|
163
|
+
if (result.confidence.overall < config.minConfidenceForAutoCitation) {
|
|
164
|
+
warnings.push({
|
|
165
|
+
code: "pdf-low-confidence",
|
|
166
|
+
message: `PDF confidence ${String(result.confidence.overall)} is below the auto-citation threshold ${String(config.minConfidenceForAutoCitation)}.`
|
|
167
|
+
});
|
|
168
|
+
}
|
|
169
|
+
return warnings;
|
|
170
|
+
}
|
|
171
|
+
function createPdfArtifacts(result, config) {
|
|
172
|
+
return [
|
|
173
|
+
{
|
|
174
|
+
kind: "generated-markdown",
|
|
175
|
+
payload: {
|
|
176
|
+
markdown: result.markdown
|
|
177
|
+
}
|
|
178
|
+
},
|
|
179
|
+
{
|
|
180
|
+
kind: "source-map",
|
|
181
|
+
payload: result.sourceMap
|
|
182
|
+
},
|
|
183
|
+
{
|
|
184
|
+
kind: "document-ir",
|
|
185
|
+
payload: result.ir
|
|
186
|
+
},
|
|
187
|
+
{
|
|
188
|
+
kind: "assets",
|
|
189
|
+
payload: result.assets
|
|
190
|
+
},
|
|
191
|
+
{
|
|
192
|
+
kind: "diagnostics",
|
|
193
|
+
payload: result.diagnostics
|
|
194
|
+
},
|
|
195
|
+
{
|
|
196
|
+
kind: "warnings",
|
|
197
|
+
payload: result.warnings
|
|
198
|
+
},
|
|
199
|
+
{
|
|
200
|
+
kind: "confidence",
|
|
201
|
+
payload: result.confidence
|
|
202
|
+
},
|
|
203
|
+
{
|
|
204
|
+
kind: "parser-options",
|
|
205
|
+
payload: createPdfConvertOptionsSummary(config)
|
|
206
|
+
}
|
|
207
|
+
];
|
|
208
|
+
}
|
|
209
|
+
function firstHeading(blocks) {
|
|
210
|
+
return blocks.find((block) => block.kind === "heading")?.text ?? null;
|
|
211
|
+
}
|
|
212
|
+
function hashString(value) {
|
|
213
|
+
return createHash("sha256").update(value).digest("hex");
|
|
214
|
+
}
|
|
215
|
+
//# sourceMappingURL=pdf2llm.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pdf2llm.js","sourceRoot":"","sources":["../../src/parsers/pdf2llm.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAEzC,OAAO,EACL,oBAAoB,EACpB,aAAa,EAMd,MAAM,gBAAgB,CAAC;AAExB,OAAO,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AAC/C,OAAO,EAAE,sBAAsB,EAAE,MAAM,kBAAkB,CAAC;AAC1D,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAS/C,MAAM,CAAC,MAAM,mBAAmB,GAAG,WAAW,CAAC;AAC/C,MAAM,CAAC,MAAM,uBAAuB,GAAG,OAAO,CAAC;AAC/C,MAAM,CAAC,MAAM,sBAAsB,GAAG,uBAAuB,CAAC;AAc9D,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,YAAoB,EACpB,OAA4B;IAE5B,MAAM,MAAM,GAAG,MAAM,kBAAkB,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC;IAC/D,MAAM,MAAM,GAAG,sBAAsB,CAAC,MAAM,CAAC,QAAQ,EAAE,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAC;IACvE,MAAM,QAAQ,GAAG;QACf,GAAG,MAAM,CAAC,QAAQ;QAClB,GAAG,gBAAgB,CAAC,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC;KAC5C,CAAC;IACF,MAAM,MAAM,GAAG,iBAAiB,CAC9B,MAAM,CAAC,MAAM,EACb,MAAM,EACN,sBAAsB,CAAC,MAAM,CAAC,QAAQ,CAAC,EACvC,QAAQ,CACT,CAAC;IACF,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAE5D,OAAO;QACL,UAAU,EAAE,mBAAmB;QAC/B,aAAa,EAAE,sBAAsB;QACrC,UAAU,EAAE,KAAK;QACjB,KAAK,EAAE,YAAY,CAAC,MAAM,CAAC;QAC3B,IAAI;QACJ,cAAc,EAAE,aAAa,CAAC,IAAI,CAAC;QACnC,MAAM;QACN,QAAQ;QACR,QAAQ,EAAE;YACR,OAAO,EAAE;gBACP,cAAc,EAAE,uBAAuB;gBACvC,aAAa;gBACb,UAAU,EAAE,MAAM,CAAC,UAAU;gBAC7B,SAAS,EAAE,MAAM,CAAC,WAAW,CAAC,KAAK,CAAC,MAAM;gBAC1C,UAAU,EAAE,MAAM,CAAC,EAAE,CAAC,UAAU;aACjC;SACF;QACD,SAAS,EAAE,kBAAkB,CAAC,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC;KACtD,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,0BAA0B,CAAC,MAAiB;IAC1D,OAAO,UAAU,CACf,IAAI,CAAC,SAAS,CAAC;QACb,MAAM,EAAE,mBAAmB;QAC3B,cAAc,EAAE,uBAAuB;QACvC,aAAa;QACb,OAAO,EAAE,8BAA8B,CAAC,MAAM,CAAC;KAChD,CAAC,CACH,CAAC;AACJ,CAAC;AAED,SAAS,kBAAkB,CACzB,YAAoB,EACpB,OAA4B;IAE5B,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,IAAI,oBAAoB,CAAC;IAC5D,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAC;IACzC,MAAM,OAAO,GAAG,UAAU,CAAC,GAAG,EAAE;QAC9B,UAAU,CAAC,KAAK,EAAE,CAAC;IACrB,CAAC,EAAE,OAAO,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;IAE7B,OAAO,SAAS,CAAC,YAAY,EAAE;QAC7B,GAAG,uBAAuB,CAAC,OAAO,CAAC,MAAM,CAAC;QAC1C,MAAM,EAAE,UAAU,CAAC,MAAM;KAC1B,CAAC;SACC,KAAK,CAAC,CAAC,KAAc,EAAE,EAAE;QACxB,IAAI,UAAU,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YAC9B,MAAM,IAAI,eAAe,CACvB,+BAA+B,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,SAAS,CAAC,QAAQ,YAAY,EAAE,EACrF,KAAK,CACN,CAAC;QACJ,CAAC;QAED,MAAM,IAAI,eAAe,CAAC,uBAAuB,YAAY,EAAE,EAAE,KAAK,CAAC,CAAC;IAC1E,CAAC,CAAC;SACD,OAAO,CAAC,GAAG,EAAE;QACZ,YAAY,CAAC,OAAO,CAAC,CAAC;IACxB,CAAC,CAAC,CAAC;AACP,CAAC;AAED,SAAS,uBAAuB,CAAC,MAAiB;IAChD,OAAO;QACL,QAAQ,EAAE;YACR,WAAW,EAAE,MAAM,CAAC,WAAW;SAChC;QACD,MAAM,EAAE;YACN,IAAI,EAAE,UAAU;SACjB;QACD,QAAQ,EAAE;YACR,SAAS,EAAE,MAAM,CAAC,SAAS;SAC5B;QACD,MAAM,EAAE;YACN,WAAW,EAAE,MAAM,CAAC,WAAW;SAChC;KACF,CAAC;AACJ,CAAC;AAED,SAAS,8BAA8B,CAAC,MAAiB;IACvD,OAAO;QACL,QAAQ,EAAE;YACR,WAAW,EAAE,MAAM,CAAC,WAAW;SAChC;QACD,MAAM,EAAE;YACN,IAAI,EAAE,UAAU;SACjB;QACD,QAAQ,EAAE;YACR,SAAS,EAAE,MAAM,CAAC,SAAS;SAC5B;QACD,MAAM,EAAE;YACN,WAAW,EAAE,MAAM,CAAC,WAAW;SAChC;QACD,QAAQ,EAAE;YACR,4BAA4B,EAAE,MAAM,CAAC,4BAA4B;SAClE;KACF,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CACxB,MAA8B,EAC9B,MAAqB,EACrB,gBAAmC,EACnC,QAAkC;IAElC,MAAM,YAAY,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAE7D,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE;QAC1B,MAAM,KAAK,GAAG,kBAAkB,CAC9B,MAAM,CAAC,SAAS,EAChB,gBAAgB,CAAC,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC,IAAI,CAAC,CAC3C,CAAC;QAEF,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;YACnB,OAAO;gBACL,GAAG,KAAK;gBACR,QAAQ,EAAE;oBACR,GAAG,CAAC,KAAK,CAAC,QAAQ,IAAI,EAAE,CAAC;oBACzB,aAAa,EAAE,MAAM,CAAC,UAAU,CAAC,OAAO;oBACxC,eAAe,EAAE,YAAY;iBAC9B;aACF,CAAC;QACJ,CAAC;QAED,OAAO;YACL,GAAG,KAAK;YACR,QAAQ,EAAE;gBACR,GAAG,CAAC,KAAK,CAAC,QAAQ,IAAI,EAAE,CAAC;gBACzB,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,SAAS,GAAG,CAAC;gBACnC,aAAa,EAAE,MAAM,CAAC,UAAU,CAAC,OAAO;gBACxC,eAAe,EAAE,YAAY;gBAC7B,gBAAgB,EAAE,KAAK,CAAC,KAAK,CAAC,IAAI;gBAClC,SAAS,EAAE;oBACT,SAAS,EAAE,KAAK,CAAC,MAAM,CAAC,SAAS;oBACjC,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;oBACjB,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;oBACjB,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,KAAK;oBACzB,MAAM,EAAE,KAAK,CAAC,MAAM,CAAC,MAAM;oBAC3B,MAAM,EAAE,KAAK,CAAC,MAAM,CAAC,MAAM;iBAC5B;aACF;SACF,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,kBAAkB,CACzB,SAA4B,EAC5B,cAAsB;IAEtB,KAAK,MAAM,KAAK,IAAI,SAAS,CAAC,OAAO,EAAE,CAAC;QACtC,IACE,cAAc,GAAG,KAAK,CAAC,aAAa;YACpC,cAAc,GAAG,KAAK,CAAC,WAAW,EAClC,CAAC;YACD,SAAS;QACX,CAAC;QAED,MAAM,MAAM,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QAEhC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;QAC3B,CAAC;IACH,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,sBAAsB,CAAC,QAAgB;IAC9C,MAAM,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC;IAEpB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACxD,IAAI,QAAQ,CAAC,KAAK,CAAC,KAAK,IAAI,EAAE,CAAC;YAC7B,OAAO,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QAC1B,CAAC;IACH,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,gBAAgB,CACvB,MAAqB,EACrB,MAAiB;IAEjB,MAAM,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;QACjD,IAAI,EAAE,OAAO,CAAC,IAAI;QAClB,OAAO,EAAE,OAAO,CAAC,OAAO;KACzB,CAAC,CAAC,CAAC;IAEJ,IAAI,MAAM,CAAC,UAAU,CAAC,OAAO,GAAG,MAAM,CAAC,4BAA4B,EAAE,CAAC;QACpE,QAAQ,CAAC,IAAI,CAAC;YACZ,IAAI,EAAE,oBAAoB;YAC1B,OAAO,EAAE,kBAAkB,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,OAAO,CAAC,yCAAyC,MAAM,CAAC,MAAM,CAAC,4BAA4B,CAAC,GAAG;SACpJ,CAAC,CAAC;IACL,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,kBAAkB,CACzB,MAAqB,EACrB,MAAiB;IAEjB,OAAO;QACL;YACE,IAAI,EAAE,oBAAoB;YAC1B,OAAO,EAAE;gBACP,QAAQ,EAAE,MAAM,CAAC,QAAQ;aAC1B;SACF;QACD;YACE,IAAI,EAAE,YAAY;YAClB,OAAO,EAAE,MAAM,CAAC,SAAS;SAC1B;QACD;YACE,IAAI,EAAE,aAAa;YACnB,OAAO,EAAE,MAAM,CAAC,EAAE;SACnB;QACD;YACE,IAAI,EAAE,QAAQ;YACd,OAAO,EAAE,MAAM,CAAC,MAAM;SACvB;QACD;YACE,IAAI,EAAE,aAAa;YACnB,OAAO,EAAE,MAAM,CAAC,WAAW;SAC5B;QACD;YACE,IAAI,EAAE,UAAU;YAChB,OAAO,EAAE,MAAM,CAAC,QAAQ;SACzB;QACD;YACE,IAAI,EAAE,YAAY;YAClB,OAAO,EAAE,MAAM,CAAC,UAAU;SAC3B;QACD;YACE,IAAI,EAAE,gBAAgB;YACtB,OAAO,EAAE,8BAA8B,CAAC,MAAM,CAAC;SAChD;KACF,CAAC;AACJ,CAAC;AAED,SAAS,YAAY,CAAC,MAA8B;IAClD,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,EAAE,IAAI,IAAI,IAAI,CAAC;AACxE,CAAC;AAED,SAAS,UAAU,CAAC,KAAa;IAC/B,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AAC1D,CAAC"}
|